View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
Hydrolysis of aspartic acid phosphoramidate nucleotides: a comparative quantum chemical study
Published on 11 June 2009. Downloaded by KU Leuven University Library on 29/09/2013 06:19:07.
Servaas Michielssens,*ab Nguyen Tien Trung,ab Matheus Froeyen,c Piet Herdewijn,c Minh Tho Nguyenab and Arnout Ceulemansab Received 25th March 2009, Accepted 18th May 2009 First published as an Advance Article on the web 11th June 2009 DOI: 10.1039/b906020k L-Aspartic acid has recently been found to be a good leaving group during HIV reverse transcriptase catalyzed incorporation of deoxyadenosine monophosphate (dAMP) in DNA. This showed that L-Asp is a good mimic for the pyrophosphate moiety of deoxyadenosine triphosphate. The present work explores the thermochemistry and mechanism for hydrolysis of several models for L-aspartic-dAMP using B3LYP/DGDZVP, MP2/6-311++G** and G3MP2 level of theory. The effect of the new compound is gradually investigated: starting from a simple methyl amine leaving group up to the aspartic acid leaving group. The enzymatic environment was mimicked by involving two Mg2+ ions and some important active site residues in the reaction. All reactions are compared to the corresponding O-coupled leaving group, which is methanol for methyl amine and malic acid for aspartic acid. With methyl amine as a leaving group a tautomeric associative or tautomeric dissociative mechanism is preferred and the barrier is lower than the comparable mechanism with methanol as a leaving group. The calculations on the aspartic acid in the enzymatic environment show that qualitatively the mechanism is the same as for triphosphate but the barrier for hydrolysis by the associative mechanism is higher for L-aspartic-dAMP than for L-malic-dAMP and pyrophosphate.
1. Introduction Triphosphate has a central role in the metabolism of living organisms. It is known to be the fuel for numerous biochemical processes.1,2 Hydrolysis of the high energy P–O bond in triphosphate is coupled to many endothermic reactions in nature. One of the processes in which triphospate plays a crucial role is DNA and RNA polymerization. Here the energy stored in triphosphate is coupled to nucleic acid polymerization. Recently Adelfinskaya et al.3,4 succeeded in finding new leaving groups replacing the pyrophosphate leaving group in triphosphate. Amino acids were coupled with a P–N bond to deoxyadenosine monophosphate (dAMP) forming phosphoramidates. Especially the results with L-histidine (L-His) and L-aspartic acid (L-Asp) were remarkable. Polymerization reactions using L-Asp-dAMP and L-His-dAMP using HIV reverse transcriptase (HIV RT) were successful in the synthesis of small strands of DNA. Best results were obtained with L-Asp-dAMP. This can be rationalized by looking at the high charge of the latter compound (3) which resembles the charge of pyrophosphate (4) better than L-His. The search for pyrophosphate alternatives in DNApolymerization reactions is important from three perspectives. a
Katholieke Universiteit Leuven, Department of Chemistry and LMCC-Mathematical Modeling and Computational Science Center, B-3001, Leuven, Belgium. E-mail:
[email protected] b Katholieke Universiteit Leuven, INPAC Institute of Nanoscale Physics & Chemistry, B-3001, Leuven, Belgium c Katholieke Universiteit Leuven, Rega Institute for Medicinal Research, Medicinal Chemistry Laboratory, B-3000, Leuven, Belgium
7274 | Phys. Chem. Chem. Phys., 2009, 11, 7274–7285
The first is in medicinal chemistry. Using nucleotides with a modified sugar chain termination in DNA polymerization can be accomplished.5 A major problem with a lot of known antiviral drugs using this principle, e.g. acyclovir6,7 or azidothymidine7 is that they need to be phosphorylated in the cell. For this they need to pass a kinase pathway which is often more selective than the viral polymerase. The phosphoramidates presented by Adelfinskaya et al.3,4 can be used to bypass this pathway. This strategy was successfully applied3 for L-Asp-PMEA (phosphomethoxyethyladenine) and sustains the high expectations of these molecules as antiviral drugs, provided a better mechanistic-based leaving group can be found. It is very important that the leaving group is a cellular metabolite. This means the leaving group is non-toxic and can be further processed by the cellular metabolism. A second application is in synthetic biology. These new building blocks for life can be further developed to take part in an orthogonal metabolism. Such orthogonal metabolism is necessary to avoid contamination of artificial life forms in the environment. Third is a fundamental understanding of polymerase reaction in nature by use of model components and more specifically to see if pyrophosphate can be replaced in this reaction. The high energy of the P–O bond in pyrophosphate has been traditionally attributed mainly to three factors.8 Firstly resonance stabilization is less favored in reactants than in products. This is claimed to be due to the ‘‘opposing resonance effect’’ caused by competition between an adjacent phosphoryl groups for the same lone pair on the bridging oxygen atom. Secondly, electrostatic repulsion is thought to be an important factor. And finally, solvation effects are shown to be less This journal is
c
the Owner Societies 2009
Published on 11 June 2009. Downloaded by KU Leuven University Library on 29/09/2013 06:19:07.
View Article Online
favorable for the reactants than for the hydrolysis products. Recently it has been suggested that the anomeric effect should replace the resonance effect in this list,8,9 this effect is shown to contribute for both P–N8 and P–O9 bonds. Mimetic compounds serving as a replacement for pyrophosphate must have similar thermodynamic properties. Here we use quantum chemical methods to calculate the enthalpy difference upon changing the P–O–CH3 bond in [MeOPO3H] to P–S or P–N and changing pyrophosphate to L-asp. We also calculate the effect of the increasing charge on hydrolysis enthalpy. Also we investigate the detailed mechanism of the dissociation path of aspartic acid in hydrolysis reactions. We mimic this reaction using 4 different models. The first is a simple model with methyl amine as leaving group. For this simple molecule we investigate different possible mechanisms. As a second model we use aspartic acid as a leaving group. Next, we examine the effect of Mg2+, since it is present in many enzymes as a catalytic metal. So in the third model we add a single Mg2+ and in the final model two Mg2+ ions. In all different models we compare what happens if the dissociating P–N bond is changed to a P–O bond. In the first model this corresponds to changing the methyl amine leaving group by methanol, this gives us an interesting opportunity to compare with recent theoretical calculations.10,11 In all other models this corresponds to changing aspartic acid to malic acid, which gives us an indication of the use of malic acid versus aspartic acid as a pyrophosphate alternative.
II.
Methods
All calculations were performed using the Gaussian 03 package.12 Visualization was done with VMD.13 The structures were optimized with the B3LYP method14 with DGDZVP15,16 basis set. Surface scans of the different molecules used were done by manually changing the dihedral angles and comparing the energies. This was done for L-Asp in different protonation states. After this mono-methyl-phosphate was added to the 8 lowest conformations and for all of them different conformations of the monophosphate were scanned. For the thermochemistry calculations G3B3 and G3MP2 methods17,18 were used. These methods have on average a small absolute deviation from experiment (0.99 kcal mol1 for G3B3 and 1.30 kcal mol1 for G3MP217,18). G3B3 theory uses geometries and zero-point energy corrections (ZPE) from B3LYP/6-31G(d) calculations after this higher level single point calculations are done. Using an interpolation procedure, energies are obtained comparable to a QCISD(T)/G3 large computation with much less computing power.19 For the reaction mechanisms the geometry optimization procedure was the same. Every stationary point was checked by frequency analysis. ZPE correction using B3LYP were scaled20 with a factor of 0.9877. The energy of the B3LYP/ DGDZVP optimized structures was corrected doing single point MP2/6-311++G** calculations on them (indicated as MP2/6-311++G**//B3LYP/DGDZVP). Single point solvation corrections using continuum solvent were performed. We used the PCM21,22 model with Pauling radii; van der Waals radii were scaled11,23–25 by 1.2. If not explicitly mentionted energies This journal is
c
the Owner Societies 2009
reported in the text are MP2 energies with solvation correction. Molar energies in tables are reported as follows: DEg is the energy in the gas phase, DGg is the free energy in the gas phase, DGsol is to solvation energy and DGaq is the free energy in aqueous solution.
III. A
Results and discussion
Influence of bond type and charge on phosphate hydrolysis
The enthalpy for different hydrolysis reactions (see Fig. 1) was calculated using a high level of theory. The first three reactions in Table 1 were calculated with both G3B3 and G3MP2,17,18 the difference between the two methods is never larger than 0.5 kcal mol1. Due to the computational complexity the last three reactions were only computed at G3MP2 level of theory. However, as shown in literature18 and confirmed for the first three reactions, the difference between the two methods is small. When analyzing nature’s energy carriers we expect P–N to be less stable than P–O comparing e.g. free energy for N 0 -phosphorylcreatine hydrolysis (10.3 kcal mol1) with ATP hydrolysis (7.3 kcal mol1)8 or phosphoester bond in phophohydroxyamino acids (6.5 to 9.5 kcal mol) with phosphoramidate in phosphohistidine(12 to 14 kcal mol1).26 This is also confirmed by the calculations. The most stable conformation used for the reactants is a conformation with the OH and CH3 group eclipsed. The hydrolysis reaction with the CH3OH leaving group is significantly less exothermic than the hydrolysis reactions with CH3NH2 and CH3SH (see Table 1). This makes the latter interesting substitutes for pyrophosphate, since reduced bonding energy might compensate for the loss of e.g. ‘‘opposing resonance stabilization’’ of L-Asp-dAMP the disadvantage is that they are less stable towards hydrolysis in solution, which implies that they could be dissociated before serving as a substrate for an enzyme. P–O bond length is 1.72 A˚, P–N 1.76 A˚ and P–S 2.20 A˚, suggesting that while CH3NH2 and CH3SH have similar hydrolysis energies CH3SH might cause a distortion in the active site of enzymes due to the high difference in bond length. To compare the results to previous results on triphosphate24,27,28 we also studied the effect of the charge on the reaction. It is well known that charge is an important factor in the exothermic nature of triphosphate hydrolysis, see e.g. Strajbl et al.29 As a model for L-Asp-dAMP10,11,30 we used the molecules
Fig. 1 Different models used for the thermochemistry calculations.
Phys. Chem. Chem. Phys., 2009, 11, 7274–7285 | 7275
View Article Online Table 1 Enthalpy of phosphate hydrolysis in kcal mol1. The reactions are presented in Fig. 1
Published on 11 June 2009. Downloaded by KU Leuven University Library on 29/09/2013 06:19:07.
XQNH XQO XQS
G3B3
G3MP2
7.7 1.7 6.4
7.7 1.6 –5.9
G3MP2 R1QOH, R2QOH R1QOH, R2QO R1QO, R2QO
1.8 58.7 127.9
in Fig. 2. Here we replace the ribose sugar and the base by hydrogen, since we are mainly interested in the pyrophosphate substitute. A similar model is used for triphosphate.24,27,28 The molecule with a charge minus two is protonated on the carboxyl function of the main chain of aspartic acid. The P–N bond length decreases with increasing charge, from 1.72 A˚ for the monoanion over 1.70 A˚ for the dianion to 1.68 A˚ for the trianion. We also applied the G3MP2 method to recently calculated triphosphate structures found in literature.27 The trend in our results is similar to previous studies, see Table 2. For the lower charged species (1 and 2) the results agree quantitatively. For the highest (3) charge there is a considerable deviation. It should also be noted that this charge effect is much smaller in solvent.28 What is important for the molecules with charge 2 and 3 is the comparison with triphosphate not the absolute value. It is clear that the trend in hydrolysis enthalpy is analogous to the trend for triphosphate. Consistent with the weaker P–N bond these values are systematically more exothermic than for triphosphate with the corresponding charge. For the 3 charged molecule, which is expected to be the state at physiological conditions, the enthalpy for phosphoramidate is 127.9 kcal mol1 and for triphosphate 96.4 kcal mol1. Overall the results suggest as expected that phosphoramidates with equal charge are more susceptible to hydrolysis than triphosphate, even in the absence of the ‘‘opposing resonance stabilization’’, which is impossible here due to the lack of competing groups for P–N conjugation. B
Reaction of [MeNHPO3H] and [MeOPO3H] + H2O
The simplest model for triphosphate studied is mono-methylphosphate ([MeOPO3H]).10,11,30 In analogy we use N-methyl-phosphoramidate ([MeNHPO3H]) as a simple model to study the effect of changing the P–O bond in [MeOPO3H] to P–N in [MeNHPO3H]. We calculated four different hydrolysis reactions and compared those to previous work10,11,30 and own calculations on the well known hydrolysis reaction of [MeOPO3H]. Before we start the discussion on the different mechanism we explain the labels used. All mechanisms are labeled as follows: X_(ts/int)m_n, the X points to the nature of the mechanism, then is indicated whether it is a transition state (ts) or an intermediate (int) and finally m labels the position on
Fig. 2 Geometries of the reactants used in the G2MP2 calculations on L-Asp-phosphoramidate [MeNHPO3H] hydrolysis.
7276 | Phys. Chem. Chem. Phys., 2009, 11, 7274–7285
the reaction pathway, and n gives the number of water molecules involved. There are four different mechanisms discussed: an associative mechanism (A), a tautomeric associative mechanism (TA), a tautomeric dissociative mechanism (TD) and interchange mechanism (Ia). The first pathway is an associative mechanism (A). In this well known mechanism water attacks the face of the tetrahedron opposite to the leaving group. As a result, a trigonal bipyramidal intermediate is formed and the configuration on the phosphorus is inverted. The P–O bond is mainly formed before the P–N bond is broken. The stationary points for this reaction are shown in Fig. 3 and energies are reported in Table 3. In the reactant complex of [MeNHPO3H] the attacking water molecule forms a double hydrogen bond with the N-methyl-phosphoramidate. The P–N bond here is 1.73 A˚. In A_ts1_1 a water molecule attacks the phosphoramidate. H1 is transferred to O3 and at the same time the O1–P1 bond (2.23 A˚) is partially formed. This TS is a strained fourmembered ring system. It is noted that an attack of a water dimer could release this ring strain and lower the barrier.31,32 We have explored this in A_ts1_2, the geometry is still strained, because of the P–O1 distance of 2.19 A˚ and the O1–P–O3 angle of 93, which are difficult to fit in a sixmembered ring. Instead of lowering, we have a slight increase of the barrier of 2.2 kcal mol1 (Table 3). As OH attacks the phosphorus a pentacoordinated trigonal bipyramidal phosphorus with the incoming and leaving group apical is formed (A_int_1). The next step is the transfer of H2 to N and the breaking of the P–N bond, this happens in A_ts2_1. The P–N bond distance is enlarged to 2.18 A˚ and the P–O1 distance reduced to 1.74 A˚. Using two water molecules in this transition state gives a remarkable result, A_ts2_1 is lowered by 7.4 kcal mol1 with the help of an extra water molecule in A_ts2_2. This was noticed for the tautomeric dissociative mechanism of [MeNHPO3H] by several authors,10,30 but not for the associative mechanism. As shown in Table 4 for the associative mechanism of [MeOPO3H] there is no reduction in the barrier if an extra water molecule is involved. The crucial parameters seem to be the length of the P–N or P–O bond. For A_ts2_2 of [MeOPO3H] the transition state is very similar (figure not shown), but P–O of the leaving group is 2.25 A˚ while for A_ts2_1 of P–N the bond distance is 2.0 A˚. This is also consistent with the fact that A_ts1_2 is not lower than A_ts1_1 and confirms the need of a relaxed six-membered ring system. The rate limiting step in the mechanism is to attain A_ts1_1, for which the barrier is 40.5 kcal mol1. Allowing for extra water molecules to participate in the reaction, the barrier for the associative mechanism of [MeOPO3H] hydrolysis is about 1.6 kcal mol1 lower than for [MeNHPO3H]. Our results at MP2/6-311**G++//B3LYP/DGDZVP level of theory and previously reported calculations at different levels of theory on [MeNHPO3H] hydrolysis differ only by about 1–2 kcal mol1 (Table 4) which validates the methods used in this work. A second mechanism is the interchange mechanism33 (Ia). The interchange mechanism as presented here is a single step mechanism with cis-attack and is to our knowledge not reported for [MeOPO3H]. Our own calculations and scans of the free energy surface in the literature11 suggest that this This journal is
c
the Owner Societies 2009
View Article Online
Table 2
Enthalpy of phosphate hydrolysis in kcal mol1. Basis set for HF,B3LYP,MP2 = 6-31++G**
Published on 11 June 2009. Downloaded by KU Leuven University Library on 29/09/2013 06:19:07.
CH4P3O10 + H2O - CH4P2O7 + H3PO4 CH4P3O102 + H2O - CH4P2O7 + H2PO4 CH4P3O103 + H2O - CH4P2O72 + H2PO4
G3B3
HF27
B3LYP27
MP227
14.5 47.9 96.4
13.7 48.5 168.4
14.5 48.8 109.3
12.6 48.5 86.0
Fig. 3 Stationary points on the pathways for the associative and interchange mechanism of [MeNHPO3H] hydrolysis.
Table 3
Relative energies for stationary point in [MeNHPO3H]-hydrolysis (in kcal mol1) B2LYP/DGDZVP DEg
DGsol
Associative 0.0 0.0 69.3 0.0 0.0 67.5 42.7 44.7 71.2 42.3 45.4 68.7 34.6 36.8 69.1 44.6 46.6 67.9 38.8 42.6 68.8 Interchange 53.4 55.3 84.0 Tautomeric associative and dissociative 28.0 28.3 69.7 14.9 16.8 75.6 7.8 7.4 77.5 41.1 40.6 74.9 11.7 11.1 62.6 6.9 3.8 56.7 31.4 20.0 63.3 25.4 24.9 67.1
React_1 React_2 A_ts1_1 A_ts1_2 A_int_1 A_ts2_1 A_ts2_2 Ia_ts1_1 TA/TD_ts1a_1 TA/TD_ts1b_1 TA/TD_int_1 TA_ts2_1 TD_ts2_1 TD_int2_1 TD_ts3_1 TA/TD_ts3_2
This journal is
DGg
MP2/6-311++G** //B3LYP/DGDZVP
c
the Owner Societies 2009
DGaq
DEg
DGg
DGsol
DGaq
0.0 0.0 42.8 44.2 36.9 48.0 41.2
0.0 0.0 40.5 40.3 35.1 43.2 37.5
0.0 0.0 42.5 43.4 37.3 45.2 41.3
72.4 70.5 74.3 71.1 71.3 71.3 71.5
0.0 0.0 40.5 42.8 38.4 47.7 40.3
40.5
54.7
56.5
89.2
39.6
27.9 10.4 0.7 35.1 17.9 16.5 36.1 25.3
27.5 14.5 12.9 42.6 16.7 13.4 35.3 28.9
27.8 16.5 12.5 42.6 16.1 10.3 33.9 28.5
73.7 80.2 81.7 77.3 65.6 60.9 66.5 70.2
26.7 8.7 3.5 37.4 23.2 22.1 40.0 28.8
Phys. Chem. Chem. Phys., 2009, 11, 7274–7285 | 7277
View Article Online
Table 4
Relative energies of [MeOPO3H] hydrolysis (in kcal mol1)
Published on 11 June 2009. Downloaded by KU Leuven University Library on 29/09/2013 06:19:07.
B3LYP/DGDZVPa DGaq React A_ts1_1 A_ts1_2 A_ts2_1 A_ts2_2 TD_ts1a_1 TD_ts1b_1 TD_ts3_1 TD_ts3_2 a
Associative 0.0 40.9 39.9 40.6 41.2 Dissociative 38.8 27.4 39.5 26.3
MP2/6-311++G**// B3LYP/DGDZVPa DGaq
B3LYP/cc-PVTZ+// B3LYP/6-31+G(d,p)b DGaq
CCSD(T)/(G2X)// B3LYP/6-31+G(d)c DGaq
38.6 39.1 39.1 38.9
36.1
37.5
37.6
38.6
36.5 24.3 38.6 24.7
35.9 24.0 32.3
33.4 23.8 33.1 23.6
This work. b Wang et al.10 c Hu et al. with DE[CCSD(T)/(G2X)] = DE[MP2/6-311+G(2df,2p)] DE[MP2/6-31 + G*] + DE[CCSD(T)/6-31 + G*].30
mechanism does not exist for [MeOPO3H]. We report it here for [MeNHPO3H]. This is a single step mechanism where P–N bond breaking and P–O1 bond formation happen at the same time (Fig. 3). Dissociation of the P–N bond is assisted by a proton transfer from water to N. The P–N bond distance is 1.98 A˚, the P–O1 distance 2.13 A˚. The H1–N bond (1.06 A˚) is almost completely formed in this step. The P–O2 bond is enlarged from 1.67 A˚ in the reactant complex to 1.77 A˚ in this transition state. The O2–P–O3, O2–P–N, O2–P–O4 and O1–P–O2 angles are, respectively 97, 83, 100 and 1531. The geometry is not a trigonal bipyramid as in the associative mechanism, where the tetragonal phosphorus is approached perpendicular to the plane formed by O3–O4–P, here the approach is not perpendicular but displaced towards the nitrogen vertex. This is caused by the hydrogen bond formed between O1 and N–H. The DGg for this interchange mechanism is 14.0 kcal mol1 higher than for the associative mechanism. But Ia_ts1_1 is stabilized by DGsol by 16.9 kcal mol1 due to the high charge separation in this TS. As a result the barrier is 0.9 kcal mol1 lower than for the associative mechanism. In a 2D plot (see Fig. 4) it is illustrated that this mechanism is
Fig. 4 Bond distances in A˚ of P–O bond broken in function of P–N bond formed along the different mechanisms. - - = associative, – = interchange, = tautomeric associative, —– = tautomeric dissociative.
7278 | Phys. Chem. Chem. Phys., 2009, 11, 7274–7285
situated between the associative and the dissociative pathway but closer to the associative. The transition states and intermediates for the third mechanism are shown in Fig. 5. This is called a tautomeric associative mechanism (TA), since this corresponds to prototropic formation of a zwitterion followed by an attack of water to the face opposite to the leaving group with inversion of configuration. The first step is a proton transfer from O2 to N with a partial rupture of the P–N bond. The P–N bond distance enlarges to 1.95 A˚. H2 is only partially transferred to N. This TS with direct proton transfer is a very strained four-membered ring system with a barrier of 26.7 kcal mol1 (Table 3). With a water molecule mediating this proton transfer, the barrier is reduced by about 18 kcal mol1 resulting in a barrier of only 8.7 kcal mol1 in TA_ts1b_1. TA_ts1b_1 is a concerted proton transfer, H2 is transferred to O1 and at the same time H3 is transferred to N. In this transition state the O2–H2 distance is 1.09 A˚ and H3–N distance 1.13 A˚ indicating that the protons are mostly situated on O2 and N. The P–N distance is 1.93 A˚ here, which is shorter than in A_ts2_2, this illustrates the importance of P–N distance for the geometry of the six-membered ring. Compared to [MeOPO3H] (Table 4) TA_ts1b_1 is about 16 kcal mol1 lower. This is consistent with the thermochemical parameters for P–N which point out that a P–O bond is stronger than a P–N bond. Further breaking of the P–N bond results in a metaphosphate, P–N bond is only partially broken and P–N distance is 2.15 A˚. DGg is 12.5 kcal mol1 for TA_int_1, this intermediate is largely stabilized by the solvent resulting in DGaq of 3.5 kcal mol1. The next rate-determining step is an attack of water to this metaphosphate. We should note here that methylamine is not completely dissociated from the metaphosphate at this point. In TA_ts2_1 the metaphosphate is attacked by water and the P–N distance is enlarged to 2.41 A˚, this is larger than in the associative and interchange mechanism, but not a complete dissociation. The O1–P distance is 2.21 A˚ which is similar to the first step in the associative mechanism. The proton H1 is transferred to one of the oxygen atoms of the metaphosphate. The role of a water dimer in this step is discussed with the tautomeric dissociative mechanism, since addition of a water dimer on TA_int1_2 results in TA/TD_ts3_2. The barriers for the tautomeric This journal is
c
the Owner Societies 2009
Published on 11 June 2009. Downloaded by KU Leuven University Library on 29/09/2013 06:19:07.
View Article Online
Fig. 5 Stationary points on the pathways of the tautomeric associative and tautomeric dissociative mechanisms of [MeNHPO3H] hydrolysis. Only metaphosphate and attacking water molecules are shown for TA/TD_ts3_2.
associative mechanism are lower than for the associative and interchange mechanism by at least 5 kcal mol1. Finally, as a fourth mechanism we have the tautomeric dissociative (TD) mechanism. Here an internal proton transfer is followed by a complete dissociation of methylamine before water attacks the metaphosphate. First transition states and intermediates are the same as in the tautomeric associative mechanism. In contrast to the TA mechanism an elongation of the P–N bond precedes the attack of a water molecule as shown in TD_ts2_1 of Fig. 5. TD_ts2_1 is a total rupture of the P–N bond (2.98 A˚) and formation of a true metaphosphate. Here we see a true dissociative mechanism, metaphosphate is now ready to be attacked by a water molecule. TD_ts3_1 shows the attack of a water monomer on the metaphosphate, here the P–N bond is completely dissociated and the P–O bond of the incoming water is shorter than in TA_ts3_1. The resulting barrier is 36.1 kcal mol1. Here again we see the importance of a water dimer in the mechanism. An attack of a water dimer on metaphosphate results in a reduced barrier. In TD_ts3_2 the barrier is reduced by 11.2 kcal mol1. Our results on the tautomeric dissociative mechanism of [MeOPO3H] hydrolysis are close to previous results (Table 4), except for TD_ts3_1 with a difference of about 6 kcal mol1, most probable due to the different solvent treatment. These results indicate that for [MeNHPO3H] the TA or TD mechanism is preferred and for [MeOPO3H] the TD mechanism, certainly when the effect of water dimers on the barriers is taken in account. The TA mechanism for [MeNHPO3H] hydrolysis is in good agreement with the mechanism proposed by Chanley et al.34 based on their experimental results. They see a complete quenching of the hydrolysis at pH higher than 10, due to deprotonation of [MeNHPO3H]. For [MeOPO3H] TD_ts1_1 has the highest barrier, while for [MeNHPO3H] this is TD_ts3_2. The largest difference in barrier for both molecules is found for TD_ts1_1, indicating that the bond breaking occurs much easier for [MeNHPO3H]. Overall we see for [MeNHPO3H] that the more we shift to a dissociative mechanism the lower the barriers get. This journal is
c
the Owner Societies 2009
Two papers exist where free energy surfaces of phosphate hydrolysis are plotted as a function of the P–O bond formed and broken.11,35 In our work no full scans were performed but the formation of the P–O bond was plotted as a function of the breaking of the P–N bond for the different stationary points (Fig. 4). For clarity, only the mechanisms involving a water monomer are plotted. This gives an overview of the nature of the pathway. The associative mechanism passes via the lower left of the graph, the tautomeric dissociative via the upper right, those are the two extreme pathways. Very close to the associative pathway we find the interchange pathway. We also see that the TD and TA mechanisms coincide in the initial phase. C
Hydrolysis reaction with aspartic acid as leaving groups
As a first model for L-Asp-dAMP3,4 hydrolysis we opted for N-(methoxyphosphinato)aspartate (Fig. 6). Similar to the model used by Bojin et al.36 we added a methyl group to one of the oxygens of the phosphate modeling the ribose and the base. The choice of this model implies that the TA and TD mechanism are excluded since the proton responsible for the internal proton transfer is replaced by a methyl. Hence only the A and Ia mechanisms are possible. The carboxyl functions were deprotonated since this is the expected state at physiological conditions. The pKa of the side chain carboxyl is 3.9 and that of the backbone carbonyl 2.1.37 The reactant complex is shown in Fig. 6. There is a hydrogen bond between the methyl group on phosphoramidate and one of the carboxyl groups. The P–N bond length is 0.02 A˚ smaller than in the reactant complex of [MeNHPO3H]. The attacking water molecule forms hydrogen bonds with the phosphoramidate and one of the carboxyls of aspartic acid. For the associative mechanism the high charge of the compound made optimization at B3LYP/DGDZVP level impossible for A_ts1_1 (Fig. 6). This structure and A_int_1 were optimized at HF/3-21G* level and then the single point energy was calculated at B3LYP/DGDZVP level. All other structures were optimized at B3LYP/DGDZVP level. This Phys. Chem. Chem. Phys., 2009, 11, 7274–7285 | 7279
Published on 11 June 2009. Downloaded by KU Leuven University Library on 29/09/2013 06:19:07.
View Article Online
Fig. 6 Stationary points on the pathways of the associative and interchange mechanism of hydrolysis reaction with aspartic acid as leaving group (charge = 3).
makes it difficult to compare the geometry of A_ts1_1 to the other calculated reactions. No reactions were calculated using extra water molecules here due to the uncertainty caused by the high charge of the system. Qualitatively this transition state is similar for [MeNHPO3H] and [MeOPO3H]. DEg is 69.8 kcal mol1 using B3LYP, a single point MP2 calculation reduces the energy to 64.4 kcal mol1. It should be stressed that this is just an estimate, we expect the true barrier to be lower. The solvation correction for this transition state is very large, this is caused by the large charge separation in this transition state, due to the proton transfer from water to phosphoramidate leaving a negatively charged OH. DGaq at B3LYP level is 57.2 kcal mol1 which is almost 15 kcal mol1 larger than for [MeNHPO3H]. Similar to [MeOPO3H]10,30 and [MeNHPO3H] this reaction proceeds through a trigonal bipyramidal intermediate. Before A_ts2_1 can occur, the H1 must rotate towards N, the transition state for this conformational change is not shown here. In A_ts2_1 a H1 is transferred from O3 to N, at the same time the P–N bond is partially broken. Again, taking into account solvent stabilization, the energy is reduced from 74.6 to 59.4 kcal mol1 at B3LYP level and from 69.6 to 53.5 kcal mol1 at MP2 level. The O1–P distance is 1.79 A˚ and the P–N distance 2.07 A˚, in [MeNHPO3H] those were, respectively 1.74 and 2.18 A˚ for [MeNHPO3H]. Also the barrier here is 12.7 kcal mol1 higher. Overall the higher charge has a large destabilizing effect on the associative mechanism. For the interchange mechanism Ia_ts1_1 (given in Fig. 6) the transfer of H1 to N and the formation of the P–O1 bond occur at the same time. The P–O1 distance is 2.23 A˚ and the P–N distance 1.93 A˚, so the formation of the P–O1 bond and breaking of the P–N bond happen at the same time. The 7280 | Phys. Chem. Chem. Phys., 2009, 11, 7274–7285
geometry of this complex has some similarities to the associative mechanism where methanol would be the leaving group. The angles O2–P–N, O2–P–O4, O2–P–O3 and O1–P–O2 are, respectively, 91, 97, 100 and 163. The O2–P bond is 1.75 A˚ and the O1–P distance 2.23 A˚, which corresponds to a highly distorted trigonal bipyramid. However, as confirmed by intrinsic reaction coordinate (IRC) calculations aspartic acid is the leaving group in this transition state and not methanol. Again the energy is reduced upon solvation from DEvac = 54.3 kcal mol1 to DGaq = 46.2 kcal mol1. The barrier for this mechanism is considerably lower than for the associative mechanism. Our results at B3LYP level suggest a reduction by 11.7 kcal mol1, at MP2 level this is reduced to 7 kcal mol1. Compared to the interchange mechanism in [MeNHPO3H] the barrier is 6.8 kcal mol1 higher. In general, we find here that an interchange mechanism is preferred and that the barrier in this highly charged compound is higher than in [MeNHPO3H]. Compared to triphosphate reactions reported in recent papers10,11 barriers reported here for L-Asp-phosphoramidate are much higher. Comparing the associative mechanism, Evac for A_ts1_1 is about 22 kcal mol1 higher and A_ts_2 is about 27 kcal mol1 higher than the values reported for the similar mechanisms at B3LYP/6-31+G(d,p) level of theory in triphosphate.10 In this paper10 solvent stabilization for triphosphate is similar to the one reported here for N-(methoxyphosphinato)aspartate. Although it is difficult to draw conclusions on this highly charged species the calculations reported here suggest that the hydrolysis reaction of L-Asp-phosphoramidate has a higher barrier than triphosphate hydrolysis and that the interchange mechanism is preferred. The high values found for the barriers seem to This journal is
c
the Owner Societies 2009
View Article Online
emphasise that these reactions are possibly unfeasible in the absence of a cationic catalyst. For this reason we introduced Mg2+.
Published on 11 June 2009. Downloaded by KU Leuven University Library on 29/09/2013 06:19:07.
D Hydrolysis reaction using Mg2+ with malic and aspartic acid as leaving groups Many enzymes involved in reactions with triphosphate have Mg2+ in their active site. We are especially interested in polymerases.38–40 Adelfinskaya et al.3 built a model of the active site of HIV-RT with L-Asp-dAMP using a known crystal in DOC structure with a dideoxynucleotide in the active site.41 We added Mg2+ to the different minima found for N-(methoxyphosphinato)aspartate and compared to the structures taken from the model of Adelfinskaya et al.3 and minimized at the same level of theory. It was found that the structure from the model of Adelfinskaya et al.3 had the lowest energy. We continued to calculate hydrolysis reaction with this structure. The reactant structure is shown in Fig. 7. The two carboxyl groups of aspartic acid, O4 and N, are coordinating Mg2+; to obtain a six-coordinated Mg2+ and complete the solvation shell38,39,41 of Mg2+ we added two water molecules similarly to Kla¨hn et al.11 The result is a distorted octahedral coordination for Mg2+. The large charge (3) in the previous model is reduced to 1 by adding Mg2+. The attacking water molecule is hydrogen bonded to the phosphate of the reactant complex this time. The P–N bond in the reactant complex is 0.03 A˚ longer here than in Fig. 6 for the structure without Mg2+. Relative energies of hydrolysis reaction are displayed in Table 5. A_ts1_1 is similar again to the previously calculated reactions. The P–O1 distance is 2.21 A˚ here and the P–N distance 1.76 A˚. Comparison to the previous section where the
same molecule was calculated without Mg2+ must be done with care since the geometry in the previous section was calculated at a lower level. We see large differences in P–N and P–O1 bond lengths in Fig. 7 but it is impossible to draw conclusions from this. Again we compared this reaction to a reaction with a water dimer. In line with the calculations on [MeNHPO3H] this is not lowering the energy. Attack of OH results again in the typical trigonal bipyramidal intermediate A_int_1. After rotation of H1 towards N we arrive at A_ts2_1, here H1 is transferred to N and the P–N bond is broken. The P–N distance here is 2.34 A˚, the P–O1 distance 1.67 A˚. Compared to the reaction without Mg2+ the P–N bond is larger and P–O1 smaller, indicating that the P–O1 bond is already largely formed and the P–N bond more dissociated. The Mg2+ is weakening the P–N bond and stabilizing the negatively charged nitrogen in this TS. The longer P–N bond in this TS suggests that an extra water molecule will not catalyze this reaction further. This is confirmed in Table 6, it is even destabilizing A_ts2_1 by 8.7 kcal mol1 at B3LYP level and by 13 kcal mol1 at MP2 level, this is in large contrast to [MeNHPO3H] where the barrier is reduced by 6.77 kcal mol1 by using a sixmembered ring. A_ts1_1 and A_ts2_1 are very close in energy at B3LYP level, the difference is larger at MP2 level, confirming that A_ts2_1 has the highest energy, as was also the case for the same system without Mg2+. The barrier for the rate limiting step is 38.8 kcal mol1 at B3LYP level and 37.4 kcal mol1 at MP2 level. Comparing with the system without Mg2+ the reduction of the barrier is spectacular, by 19.8 kcal mol1 at B3LYP level and by 12.4 kcal mol1 at MP2 level. Still the barrier is about 8 kcal mol1 higher than for a similar model with diphosphate.11 The associative mechanism was also calculated for a similar system with malic acid instead of aspartic acid as leaving
Fig. 7 Stationary points on the pathways of the associative and interchange mechanism of hydrolysis reaction using Mg2+ with aspartic acid as leaving group (charge of the hydrolysed molecule = 3, charge of total system = 1).
This journal is
c
the Owner Societies 2009
Phys. Chem. Chem. Phys., 2009, 11, 7274–7285 | 7281
View Article Online
Table 5
Relative energies of hydrolysis reaction with aspartic acid as leaving group (charge = 3) (in kcal mol1)
Published on 11 June 2009. Downloaded by KU Leuven University Library on 29/09/2013 06:19:07.
B3LYP/DGDZVP
React A_ts1_1a A_int_1a A_ts2_1 Ia_ts1_1 a
MP2/6-311++G** //B3LYP/DGDZVP
DEg
DGg
DGsol
DGaq
DEg
DGg
DGsol
DGaq
0.0 69.8 64.4 71.4 54.6
0.0 70.8 69.5 74.6 55.5
368.2 385.5 385.8 381.0 375.8
0.0 57.2 51.9 59.4 47.7
0.0 64.4 58.2 65.8 54.3
0.0 69.4 63.3 69.6 55.2
372.2 387.6 387.7 384.7 380.6
0.0 53.2 48.0 53.5 46.2
This structure was only optimized at HF/3-21G* level, reported energies are single point energies.
Table 6 Relative energies for hydrolysis reaction using Mg2+ with aspartic acid as leaving group (charge of the hydrolysed molecule = 3, charge of total system = 1) (in kcal mol1) B3LYP/DGDZVP
React_1 React_2 A_ts1_1 A_ts1_2 A_int_1 A_ts2_1 A_ts2_2 Ia_ts1_1
MP2/6-311++G** //B3LYP/DGDZVP
DEg
DGg
DGsol
DGaq
DEg
DGg
DGsol
DGaq
0.0 0.0 35.1 35.6 24.4 36.4 45.0 60.6
0.0 0.0 37.4 38.5 26.8 39.6 46.6 61.6
65.1 67.1 64.1 65.9 65.0 64.4 65.5 76.6
0.0 0.0 38.4 39.7 26.8 39.6 48.3 50.9
0.0 0.0 33.6 33.9 25.0 36.5 44.3 63.4
0.0 0.0 35.9 36.8 27.3 38.9 46.0 64.5
72.6 71.7 71.2 71.2 71.7 70.5 65.5 86.1
0.0 0.0 37.4 38.8 28.2 41.1 54.2 51.7
group. This is largely the same system except that the leaving group is coupled with a P–O bond rather than P–N. There is only one TS along this pathway, the lowest barrier here is using two water molecules (A_ts1_2), for this TS the barrier is 36.2 kcal mol1. The A_ts2 transition state doesn’t exist here: upon rotation of H1 the proton is directly transferred to the oxygen of the leaving group. The difference in barrier between the N coupled systems is 1 kcal mol1 at B3LYP level and 1.1 kcal mol1 at MP2 level. As also seen for [MeOPO3H] and [MeNHPO3H], for the associative mechanism O-coupled leaving groups have lower barriers than N-coupled leaving groups. The barrier here is closer to the barrier of 32 kcal mol1 estimated for diphosphate by Kla¨hn et al.11 In Fig. 7 Ia_ts1_1 for the reaction with aspartic acid is shown. This is the transition state for a single step mechanism. P–N distance is 1.94 A˚ and P–O1 distance is 2.17 A˚. Both are shorter than in the system without Mg2+. The geometry around the phosphorus is even more distorted from a trigonal bipyramid: O2–P–O4, O3–P–O4, N–P–O4 and O1–P–O4 angles are, respectively 108.68, 100.03, 93.7 and 160.2. For the proton transfer from O1 to nitrogen to occur an inversion at the nitrogen is necessary. This explains the higher barrier for this mechanism compared to the mechanism without Mg2+. Using B3LYP the barrier with Mg2+ is about 3 kcal mol1 higher, with MP2 it is 5 kcal mol1 higher than in the absence of Mg2+. The barrier is around 12.1 kcal mol1 higher than the equivalent barrier for [MeNHPO3H]. In the presence of a single Mg2+, the barrier for the interchange mechanisms is considerably higher than that for the associative mechanism, by 11.3 kcal mol1 at B3LYP level and 12.9 kcal mol1 at MP2 level. We conclude that the reaction with Mg2+ occurs preferentially by an associative mechanism. This is also the mechanism proposed by Bojin et al.36 and Alberts et al.40 for the polymerization reaction of polymerase b. 7282 | Phys. Chem. Chem. Phys., 2009, 11, 7274–7285
E Hydrolysis reaction using two Mg2+ with aspartic and malic acids as leaving groups As a final model we involve a second Mg2+ ion in the reaction (Fig. 8). This second metal ion, the catalytic Mg2+ is an electrophilic catalyst that stabilizes the charge of the hydroxide ion attacking the phosphorus. Polymerases often make use of those two Mg2+ ions.39,40,42 The relative energies for hydrolysis reaction using Mg2+ with malic acid as leaving group are given in Table 7. We again started from the model built in the work of Adelfinskaya et al.3 We kept the groups coordinating Mg2+ and added the attacking water molecule. The reacting molecules are the same as before, an overview is given in Fig. 9. The catalytic Mg2+ cat is coordinated by five ligands: an oxygen of asp110, of asp185, of asp186, of phosphate and of water. Different from some previous models we did not include an extra water molecule to obtain six-coordinated Mg2+. The aspartic acid coordinating Mg2+ asp is linked to six ligands: oxygen of val111, of asp110, of asp185, of the carboxyl groups of aspartic acid and of phosphate. Different from the previous model but similar to the model of Adelfinskaya et al.3 the N is not in the first coordination sphere of the Mg2+. This model is similar to models proposed for DNA-polymerase b36,43 or HIV-RT.38 In Table 8 we reported the distances for the coordination of Mg2+ and compared them to the X-ray structure41 and the model by Adelfinskaya et al.3 All structures in this section are optimized at B3LYP/6-31G level. Remarkable is the Mgcat-OD1(Asp186), Asp186 migrates closer to Mgcat by 1.84 A˚, this was also the case in the simulations of Adelfinskaya et al.3 and in some models of Rungrotmongkol et al.39 Overall there are no large differences in distance between the averages of simulations of Adelfinskaya et al.3 and the active site optimization done here. This journal is
c
the Owner Societies 2009
Published on 11 June 2009. Downloaded by KU Leuven University Library on 29/09/2013 06:19:07.
View Article Online
Fig. 8 Stationary points on the pathways of the associative mechanism of hydrolysis reaction using two Mg2+ with aspartic acid as leaving group (charge of the hydrolysed molecule = 3, charge of total system = 2). Table 7 Relative energies for hydrolysis reaction using Mg2+ with malic acid as leaving group (charge of the hydrolysed molecule = 3, charge of total system = 1) (in kcal mol1) B3LYP/DGDZVP
React_1 React_2 A_ts1_1 A_ts1_2 Ia_ts1_1
MP2/6-311++G** //B3LYP/DGDZVP
DEg
DGg
DGsol
DGaq
DEg
DGg
DGsol
DGaq
0.0 0.0 34.0 34.4 43.9
0.0 0.0 36.5 37.2 45.9
66.3 65.9 65.4 66.5 67.5
0.0 0.0 37.4 37.2 45.7
0.0 0.0 32.2 32.0 46.2
0.0 0.0 34.7 34.9 48.2
73.6 71.7 72.0 72.6 75.3
0.0 0.0 36.3 35.5 47.6
Fig. 9 Model of the active site of HIV-RT with N-(methoxyphosphinato)aspartate.
Only the associative mechanism was calculated here, several other studies on similar systems point out that an associative mechanism is preferred, 36,40 moreover the geometry of the HIV-RT active site makes an interchange mechanism impossible. As before, the highest barrier is the deprotonation step with a concerted attack of the hydroxide ion to phosphorus in A_ts1_1. The Mg2+ O1 distance decreases by 0.07 A˚ to This journal is
c
the Owner Societies 2009
2.00 A˚. We see here that involving the Mg2+ cat lowers this barrier by 8.9 kcal mol1. The relative energies are given in Table 9. Our model is very similar to the model of Bojin et al.36 for polymerase b. An important difference is that in their model the phosphorus is attacked by methanol and the leaving group is pyrophosphate. The value they computed for the first intermediate is 28 kcal mol1 at B3LYP/6-31G(d,p) level compared to our value for aspartate of 29.5 kcal mol1 on B3LYP/DGDZVP level. This indicates that the barrier for triphosphate hydrolysis will be slightly lower than for aspartic acid. No transition states were optimized in the work of Bojin et al., they estimated that the barrier was 15 kcal mol1 higher than the first intermediate, which is much too high compared to the experimental value of 16 kcal mol1 in polymerase b. Finally if we compare to the reaction with malic acid as a leaving group (see Table 10) we observe that the barrier is 1.5 kcal mol1 lower, indicating that malic acid might be a better alternative for pyrophosphate. The reaction proceeds similarly to previous reaction with one Mg2+. A_ts2_1 is higher by 2.3 kcal mol1, which might be expected since N is not linked to Mg2+ as in the previous model.
IV.
Conclusion
Recalculation of triphosphate hydrolysis reactions at higher level of theory and comparison to hydrolysis of phosphoramidates with aspartic acid as leaving group confirms that replacing Phys. Chem. Chem. Phys., 2009, 11, 7274–7285 | 7283
View Article Online Table 8 Active site distances for crystal structure,41 simulations with the amber99 force field of Adelfinskaya et al.3 and QM optimization (B3LYP/6-31G), in A˚
Published on 11 June 2009. Downloaded by KU Leuven University Library on 29/09/2013 06:19:07.
X-Ray41 Mgcat–H2O Mgcat–O3 0 (primer terminus) Mgcat–OD2(D110) Mgcat–OD1(D185) Mgcat–OD1(D186) Mgcat–O4 (P) Mgasp–OD1(D110) Mgasp–O(V111) Mgasp–OD1(D185) Mgasp–O4 (P) Mgasp–OD2 (Asp) Mgasp–OD3 (Asp) H2O–P O3 0 (primer terminus)-P Mgcat–Mgasp
Adelfinskaya et al.3
QM optimization 2.07
2.09 1.90 1.88 1.83 2.38 1.92 2.06 1.88 2.07 1.89 1.95
2.68 2.05 3.87 3.03 2.13 2.27 2.36 2.21
2.01 2.04 2.03 2.09 2.12 2.18 2.06 2.13 2.04 2.12 3.25
3.09 3.70
3.57
3.53
Table 9 Relative energies for hydrolysis reaction using two Mg2+ with aspartic acid as leaving group (in kcal mol1), (charge of the hydrolysed molecule =3, charge of total system = 2) B3LYP/6-31G
React A_ts1_1 A_int_1 A_int_1 A_ts2_1
B3LYP/DGDZVP //B3LYP/631G
DEg
DGg
DGsol
DGaq
DEg
DGg
DGsol
DGaq
0.0 21.4 20.6 22.8 32.1
0.0 22.0 21.0 23.6 32.0
161.1 161.0 161.4 164.1 163.1
0.0 22.2 20.7 20.6 31.0
0.0 30.0 30.1 31.4 43.7
0.0 30.6 30.5 32.2 44.6
152.4 153.4 156.8 156.0 155.0
0.0 29.5 29.1 28.5 42.0
Table 10 Relative energies for hydrolysis reaction using two Mg2+ with malic acid as leaving group (charge of the hydrolysed molecule = 3, charge of total system = 2), (in kcal mol1) B3LYP/6-31G
React A_ts1_1 A_int_1 A_int_1 A_ts2_1
B3LYP/DGDZVP //B3LYP/6-31G
DEg
DGg
DGsol
DGaq
DEg
DGg
DGsol
DGaq
0.0 21.7 19.0 16.2 29.0
0.0 22.2 19.1 16.8 29.6
159.8 160.5 162.9 162.5 163.1
0.0 21.4 16.0 14.0 28.2
0.0 29.0 28.7 27.6 43.3
0.0 29.5 28.8 28.2 43.8
150.8 152.4 154.5 154.8 153.6
0.0 28.0 25.1 24.2 41.1
pyrophosphate by aspartic acid results in higher hydrolysis enthalpies. Also the trend when increasing the charge is similar. This indicates that from a point of view of thermodynamics aspartic acid can serve as a good replacement of pyrophosphate where the P–N bond in phosphoramidate is even less stable than the high energy P–O bond in triphosphate. From calculations on the simplest systems [MeNHPO3H] and [MeOPO3H] we can draw three important conclusions. The first is that the use of a water dimer plays a crucial role in lowering the barrier for the tautomeric dissociative mechanism and tautomeric associative mechanism, while it does not reduce the barrier significantly for the other mechanisms. The second conclusion concerns the hydrolysis mechanisms, we see that two extra mechanisms are possible when we compare [MeNHPO3H] to [MeOPO3H], the interchange mechanism and the tautomeric associative mechanism. We also see a strong preference for a tautomeric mechanism where a zwitterionic intermediate is formed for [MeNHPO3H] hydrolysis. A third conclusion arises from comparing [MeNHPO3H] to [MeOPO3H] for both a tautomeric 7284 | Phys. Chem. Chem. Phys., 2009, 11, 7274–7285
mechanism if preferred. Comparing the associative mechanisms the barriers are 40.5 and 38.6 kcal mol1 for [MeNHPO3H] and [MeOPO3H], respectively. The pathway to achieving a metaphosphate is much lower for [MeNHPO3H] than for [MeOPO3H] this might indicate that phosphoramidates are better substrates for enzymes using a dissociative mechanism.44 While in enzymes using an associative mechanism (which is expected for HIV-RT with normal nucleotides) the barrier for pyrophosphate is lower than for L-Asp-NMP. This is also seen in the experiments by Adelfinskaya et al.3 Subsequently aspartic acid was compared as a leaving group to previous work on triphosphate. Barriers for this highly charged compound were much higher than for [MeNHPO3H] and for triphosphate. The choice of the model excluded the possibility of a dissociative mechanism and because of the high charge it was found difficult to optimize TS. Here the interchange mechanism is preferred but the barrier is much higher than for triphosphate, emphasising the need for cationic agents. This journal is
c
the Owner Societies 2009
Published on 11 June 2009. Downloaded by KU Leuven University Library on 29/09/2013 06:19:07.
View Article Online
For this reason we introduced Mg2+. Barriers for the associative mechanism were spectacularly reduced compared to the model without Mg2+ and slightly compared to [MeNHPO3H]. Here barriers are still higher than the estimate for triphosphate with one Mg2+ of Kla¨hn et al.11 and can be reduced by about 1 kcal mol1 using malic acid as leaving group. Finally inclusion of two Mg2+ ions in the reaction using an active site model reduces the barrier to 29.52 kcal mol1. Compared to the system with one Mg2+ this is a reduction of 8.9 kcal mol1. This demonstrates the crucial role of the second Mg2+ ion. The barrier for N-(methoxyphosphinato)aspartate is always higher in energy than for triphosphate. It can be reduced by the use of malic acid. Comparison of our intermediates to the intermediates in the model of Bojin et al.36 suggests that with malic acid as a leaving group the barrier is similar to triphosphate while with aspartic acid the barrier is higher. The present QM results also leave little doubt that the proper barrier for hydrolysis is much higher than the actual barrier in the enzymatic reaction. We suspect that the protein environment plays a crucial role in promoting fast proton transfer. A further detailed hybrid QM/MM study will be needed to elucidate the catalytic role of the protein. We can conclude from comparison of this study with previous results in the literature on triphosphate hydrolysis that barriers are comparable but slightly higher and that malic acid might be an improvement over aspartic acid as a leaving group.
13 14 15 16 17 18 19 20 21 22 23 24 25
Acknowledgements This work has been supported by the Fund for Scientific Research-Flanders (FWO). Janne Michielssens is acknowledged for carefully reading the paper.
26 27 28 29
References 1 H. Lodisch, A. Berk, L. S. Zipursky, P. Matsudaira, D. Baltimore and J. Darnell, Molecular Cell Biology, Freeman, 4th edn, 2000, p. 326. 2 R. H. Garrett and C. M. Grisham, Biochemistry, Saunders College Pub, 1999, p. 62. 3 O. Adelfinskaya, M. Terrazas, M. Froeyen, P. Marliere, K. Nauwelaerts and P. Herdewijn, Nucleic Acid Res., 2007, 35, 5060–5072. 4 O. Adelfinskaya and P. Herdewijn, Angew. Chem., Int. Ed., 2007, 46, 4356–4358. 5 E. J. Arts, J. P. Marois, Z. X. Gu, S. F. J. LeGrice and M. A. Wainberg, J. Virol., 1996, 70, 712–720. 6 D. D. Ilsley, S. H. Lee, W. H. Miller and R. D. kuchta, Biochemistry, 1995, 34, 2504–2510. 7 E. De Clercq, Nat. Rev. Drug Discov., 2002, 1, 13–25. 8 E. A. Ruben, M. S. Chapman and J. D. Evanseck, J. Am. Chem. Soc., 2005, 127, 17789–17798. 9 E. A. Ruben, J. A. Plumley, M. S. Chapman and J. D. Evanseck, J. Am. Chem. Soc., 2008, 130, 3349–3358. 10 Y. N. Wang, I. A. Topol, J. R. Collins and S. K. Burt, J. Am. Chem. Soc., 2003, 125, 13265–13273. 11 M. Kla¨hn, E. Rosta and A. Warshel, J. Am. Chem. Soc., 2006, 128, 15310–15323. 12 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. J. A. Montgomery, T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
This journal is
c
the Owner Societies 2009
30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez and J. A. Pople, GAUSSIAN 0.3 (Revision D.02), Gaussian Inc.,Wallingford, CT, 2004. W. Humphrey, A. Dalke and K. Schulten, J. Mol. Graph., 1996, 14, 33–39. A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652. N. Godbout, D. R. Salahub, J. Andzelm and E. Wimmer, Can. J. Chem., Rev. Can. Chim., 1992, 70, 560–571. C. Sosa, J. Andzelm, B. C. Elkin, E. Wimmer, K. D. Dobbs and D. A. Dixon, J. Phys. Chem., 1992, 96, 6630–6636. L. A. Curtiss, P. C. Redfern, K. Raghavachari, V. Rassolov and J. A. Pople, J. Chem. Phys., 1999, 110, 4703–4709. A. G. Baboul, L. A. Curtiss, P. C. Redfern and K. Raghavachari, J. Chem. Phys., 1999, 110, 7650–7. I. N. Levine, Quantum Chemistry, Prentice Hall, New Jersey, 5th edn, 2000, pp. 592–593. M. P. Andersson and P. Uvdal, J. Phys. Chem. A, 2005, 109, 2937–2941. S. Miertus, E. Scrocco and J. Tomasi, Chem. Phys., 1981, 55, 117–129. B. Mennucci and J. Tomasi, J. Chem. Phys., 1997, 106, 5151–5158. J. Florian and A. Warshel, J. Phys. Chem. B, 1998, 102, 719–734. B. Y. Ma, C. Meredith and H. F. Schaeffer, J. Phys. Chem., 1994, 98, 8216–8223. M. Bianciotto, J. C. Barthelat and A. Vigroux, J. Phys. Chem. A, 2002, 106, 6521–6526. P. V. Attwood, M. J. Piggott, X. L. Zu and P. G. Besant, Amino Acids, 2007, 32, 145–156. P. Hansia, N. Guruprasad and S. Vishveshwara, Biophys. Chem., 2006, 119, 127–136. M. E. Colvin, E. Evleth and Y. Akacem, J. Am. Chem. Soc., 1995, 117, 4357–4362. M. Strajbl, A. Shurki and A. Warshel, Proc. Natl. Acad. Sci. U. S. A., 2003, 100, 14834–14839. C. H. Hu and T. Brinck, J. Phys. Chem. A, 1999, 103, 5379–5386. M. T. Nguyen, M. H. Matus, V. E. Jackson, V. T. Ngan, J. R. Rustad and D. A. Dixon, J. Phys. Chem. A, 2008, 112, 10386–10398. M. T. Nguyen, G. Raspoet, L. G. Vanquickenborne and P. T. VanDuijnen, J. Phys. Chem. A, 1997, 101, 7379–7388. D. Katakis and G. Gordon, Mechanisms of Inorganic Reactions, John Wiley and Sons, 1st edn, 1987, pp. 170–174. J. D. Chanley and E. Feageson, J. Am. Chem. Soc., 1963, 85, 1181–1190. S. C. L. Kamerlin and J. Wilkie, Org. Biomol. Chem., 2007, 5, 2098–2108. M. D. Bojin and T. Schlick, J. Phys. Chem. B, 2007, 111, 11244–11252. R. H. Garrett and C. M. Grisham, Biochemistry, Saunders College Pub, 1999, p. 84. T. Rungrotmongkol, S. Hannongbua and A. Mulholland, J. Theor. Comput. Chem., 2004, 3, 491–500. T. Rungrotmongkol, A. J. Mulholland and S. Hannongbua, J. Mol. Graph., 2007, 26, 1–13. I. L. Alberts, Y. Wang and T. Schlick, J. Am. Chem. Soc., 2007, 129, 11100–11110. H. F. Huang, R. Chopra, G. L. Verdine and S. C. Harrison, Science, 1998, 282, 1669–1675. L. S. Beese and T. A. Steitz, EMBO J., 1991, 10, 25–33. R. C. Rittenhouse, W. K. Apostoluk, J. H. Miller and T. P. Straatsma, Proteins, 2003, 53, 667–682. W. R. Wang, H. S. Cho, R. Kim, J. Jancarik, H. Yokota, H. H. Nguyen, I. V. Grigoriev, D. E. Wemmer and S. H. Kim, J. Mol. Biol., 2002, 319, 421–431.
Phys. Chem. Chem. Phys., 2009, 11, 7274–7285 | 7285
